Among the many HST images which APOD has used, more than once, this one of Abell 2218 (taken Jan. 11 to 13, 2000 by the veteran WFPC2) has featured thrice, and its predecessor twice.

A quick count turns up some ~20 APODs showing (strong) gravitational lensing of (background) objects by (foreground) clusters (some are repeats).

Gravitational lensing is an example of the bending of light by mass, predicted by Einstein's GR, and its first observation - during a solar eclipse (stars 'near' the Sun were just where they were predicted to be*) - received widespread publicity.

These "luminous arcs" were first written up by Roger Lynds and Vahé Petrosian, in 1989 (though they were announced three years earlier).

Under favourable circumstances, the distorted images of the background object(s) can be used to reconstruct the 'gravitational lens' - if you assume light bends per Einstein's GR, from you can 'work backwards' to make a 'map' of the distribution of the mass doing the lensing.

CL0024+1654 is a good example. This page gives a popular account of how the reconstruction works (I've seen a better account, but can't find it just now); here is the preprint of the paper giving the full details (it was published, but you need a subscription; I don't think there's any significant difference between the preprint and the published paper).

This is, of course, both an independent corroboration of the 'Zwicky virial theorem' observations (covered in post #24), as well as an advance - for at least some rich clusters, a detailed 'mass map' can be made**.

For the purposes of this thread, the result is the same - most of the mass in these rich clusters is not in the galaxies, but in the space between them; and it is more concentrated towards the centre than the edges.

Clearly, there is a lot more mass in these clusters than meets the 'optical' eye.

Are there other ways to 'weigh' such clusters?

And how to tell whether the invisible stuff is made of baryons?

Stay tuned!

*The data were, in fact, quite marginal; the error bars much too large to rule out 'Newtonian' bending. Subsequent observations, in many wavebands, of bending by many objects, have all been consistent with GR, and (so far) no alternatives have come even close to matching the (good) data.
**Searching with ADS or Google Scholar, for the rich clusters featured in APODs (such as Abell 2218, Abell 1689, and CL0024+1654) is fun - there are hundreds of papers on these objects, many on reconstructing mass maps!